Additional review fixes + uriparser sync with latest master

ext/uri/php_uri.c

@@ -53,6 +53,7 @@ static PHP_MINFO_FUNCTION(uri)
 {
 	php_info_print_table_start();
 	php_info_print_table_row(2, "uri support", "active");
+	php_info_print_table_row(2, "uriparser library version", URIPARSER_VERSION);
 	php_info_print_table_end();
 }
 
@@ -64,9 +65,6 @@ static PHP_MSHUTDOWN_FUNCTION(uri)
 
 PHP_RINIT_FUNCTION(uri)
 {
-#if defined(COMPILE_DL_URI) && defined(ZTS)
-	ZEND_TSRMLS_CACHE_UPDATE();
-#endif
 
 	return SUCCESS;
 }
@@ -89,10 +87,3 @@ zend_module_entry uri_module_entry = {
 	PHP_VERSION,                    /* Version */
 	STANDARD_MODULE_PROPERTIES
 };
-
-#ifdef COMPILE_DL_URI
-#ifdef ZTS
-ZEND_TSRMLS_CACHE_DEFINE()
-#endif
-ZEND_GET_MODULE(uri)
-#endif

ext/uri/php_uri.h

@@ -20,8 +20,5 @@
 extern zend_module_entry uri_module_entry;
 #define phpext_uri_ptr &uri_module_entry
 
-#if defined(ZTS) && defined(COMPILE_DL_URI)
-ZEND_TSRMLS_CACHE_EXTERN()
-#endif
 
 #endif

ext/uri/uriparser/include/uriparser/Uri.h

@@ -85,7 +85,7 @@ extern "C" {
 # include "UriBase.h"
 #endif
 
-extern const URI_CHAR * const URI_FUNC(SafeToPointTo);
+
 
 /**
  * Specifies a range of characters within a string.
@@ -200,11 +200,7 @@ typedef struct URI_TYPE(QueryListStruct) {
 	struct URI_TYPE(QueryListStruct) * next; /**< Pointer to the next key/value pair in the list, can be NULL if last already */
 } URI_TYPE(QueryList); /**< @copydoc UriQueryListStructA */
 
-URI_PUBLIC UriBool URI_FUNC(IsHostSet)(const URI_TYPE(Uri) * uri);
 
-URI_PUBLIC UriBool URI_FUNC(PushPathSegment)(URI_TYPE(ParserState) * state,
-		const URI_CHAR * first, const URI_CHAR * afterLast,
-		UriMemoryManager * memory);
 
 /**
  * Parses a RFC 3986 %URI.
@@ -354,7 +350,7 @@ URI_PUBLIC int URI_FUNC(FreeUriMembersMm)(URI_TYPE(Uri) * uri,
 
 
 /**
- * Percent-encodes all unreserved characters from the input string and
+ * Percent-encodes all but unreserved characters from the input string and
  * writes the encoded version to the output string.
  *
  * NOTE: Be sure to allocate <b>3 times</b> the space of the input buffer for
@@ -388,7 +384,7 @@ URI_PUBLIC URI_CHAR * URI_FUNC(EscapeEx)(const URI_CHAR * inFirst,
 
 
 /**
- * Percent-encodes all unreserved characters from the input string and
+ * Percent-encodes all but unreserved characters from the input string and
  * writes the encoded version to the output string.
  *
  * NOTE: Be sure to allocate <b>3 times</b> the space of the input buffer for
@@ -465,7 +461,7 @@ URI_PUBLIC const URI_CHAR * URI_FUNC(UnescapeInPlace)(URI_CHAR * inout);
 
 /**
  * Performs reference resolution as described in
- * <a href="http://tools.ietf.org/html/rfc3986#section-5.2.2">section 5.2.2 of RFC 3986</a>.
+ * <a href="https://datatracker.ietf.org/doc/html/rfc3986#section-5.2.2">section 5.2.2 of RFC 3986</a>.
  * Uses default libc-based memory manager.
  * NOTE: On success you have to call uriFreeUriMembersA on \p absoluteDest manually later.
  *
@@ -488,7 +484,7 @@ URI_PUBLIC int URI_FUNC(AddBaseUri)(URI_TYPE(Uri) * absoluteDest,
 
 /**
  * Performs reference resolution as described in
- * <a href="http://tools.ietf.org/html/rfc3986#section-5.2.2">section 5.2.2 of RFC 3986</a>.
+ * <a href="https://datatracker.ietf.org/doc/html/rfc3986#section-5.2.2">section 5.2.2 of RFC 3986</a>.
  * Uses default libc-based memory manager.
  * NOTE: On success you have to call uriFreeUriMembersA on \p absoluteDest manually later.
  *
@@ -512,7 +508,7 @@ URI_PUBLIC int URI_FUNC(AddBaseUriEx)(URI_TYPE(Uri) * absoluteDest,
 
 /**
  * Performs reference resolution as described in
- * <a href="http://tools.ietf.org/html/rfc3986#section-5.2.2">section 5.2.2 of RFC 3986</a>.
+ * <a href="https://datatracker.ietf.org/doc/html/rfc3986#section-5.2.2">section 5.2.2 of RFC 3986</a>.
  * NOTE: On success you have to call uriFreeUriMembersMmA on \p absoluteDest manually later.
  *
  * @param absoluteDest     <b>OUT</b>: Result %URI
@@ -628,10 +624,10 @@ URI_PUBLIC int URI_FUNC(ToStringCharsRequired)(const URI_TYPE(Uri) * uri,
 
 /**
  * Converts a %URI structure back to text as described in
- * <a href="http://tools.ietf.org/html/rfc3986#section-5.3">section 5.3 of RFC 3986</a>.
+ * <a href="https://datatracker.ietf.org/doc/html/rfc3986#section-5.3">section 5.3 of RFC 3986</a>.
  *
  * NOTE: Scheme-based normalization
- * (<a href="http://tools.ietf.org/html/rfc3986#section-6.2.3">section 6.2.3 of RFC 3986</a>)
+ * (<a href="https://datatracker.ietf.org/doc/html/rfc3986#section-6.2.3">section 6.2.3 of RFC 3986</a>)
  * is not applied and is considered a responsibility of the application using uriparser.
  *
  * @param dest           <b>OUT</b>: Output destination

ext/uri/uriparser/src/UriCommon.c

@@ -431,8 +431,8 @@ unsigned char URI_FUNC(HexdigToInt)(URI_CHAR hexdig) {
 
 
 URI_CHAR URI_FUNC(HexToLetter)(unsigned int value) {
-	/* Uppercase recommended in section 2.1. of RFC 3986 *
-	 * http://tools.ietf.org/html/rfc3986#section-2.1    */
+	/* Uppercase recommended in section 2.1. of RFC 3986         *
+	 * https://datatracker.ietf.org/doc/html/rfc3986#section-2.1 */
 	return URI_FUNC(HexToLetterEx)(value, URI_TRUE);
 }
 

ext/uri/uriparser/src/UriCommon.h

@@ -70,6 +70,7 @@
 
 /* Used to point to from empty path segments.
  * X.first and X.afterLast must be the same non-NULL value then. */
+extern const URI_CHAR * const URI_FUNC(SafeToPointTo);
 extern const URI_CHAR * const URI_FUNC(ConstPwd);
 extern const URI_CHAR * const URI_FUNC(ConstParent);
 
@@ -90,6 +91,8 @@ unsigned char URI_FUNC(HexdigToInt)(URI_CHAR hexdig);
 URI_CHAR URI_FUNC(HexToLetter)(unsigned int value);
 URI_CHAR URI_FUNC(HexToLetterEx)(unsigned int value, UriBool uppercase);
 
+UriBool URI_FUNC(IsHostSet)(const URI_TYPE(Uri) * uri);
+
 UriBool URI_FUNC(CopyPath)(URI_TYPE(Uri) * dest, const URI_TYPE(Uri) * source,
 		UriMemoryManager * memory);
 UriBool URI_FUNC(CopyAuthority)(URI_TYPE(Uri) * dest,

ext/uri/uriparser/src/UriNormalize.c

@@ -104,7 +104,7 @@ static UriBool URI_FUNC(ContainsUglyPercentEncoding)(const URI_CHAR * first,
 
 static void URI_FUNC(LowercaseInplace)(const URI_CHAR * first,
 		const URI_CHAR * afterLast);
-static void URI_FUNC(LowercaseInplaceAfterPercentEncoding)(const URI_CHAR * first,
+static void URI_FUNC(LowercaseInplaceExceptPercentEncoding)(const URI_CHAR * first,
 		const URI_CHAR * afterLast);
 static UriBool URI_FUNC(LowercaseMalloc)(const URI_CHAR ** first,
 		const URI_CHAR ** afterLast, UriMemoryManager * memory);
@@ -247,7 +247,7 @@ static URI_INLINE void URI_FUNC(LowercaseInplace)(const URI_CHAR * first,
 
 
 
-static URI_INLINE void URI_FUNC(LowercaseInplaceAfterPercentEncoding)(const URI_CHAR * first,
+static URI_INLINE void URI_FUNC(LowercaseInplaceExceptPercentEncoding)(const URI_CHAR * first,
 		const URI_CHAR * afterLast) {
 	if ((first != NULL) && (afterLast != NULL) && (afterLast > first)) {
 		URI_CHAR * i = (URI_CHAR *)first;
@@ -256,6 +256,9 @@ static URI_INLINE void URI_FUNC(LowercaseInplaceAfterPercentEncoding)(const URI_
 			if ((*i >= _UT('A')) && (*i <=_UT('Z'))) {
 				*i = (URI_CHAR)(*i + lowerUpperDiff);
 			} else if (*i == _UT('%')) {
+				if (i + 3 >= afterLast) {
+					return;
+				}
 				i += 2;
 			}
 		}
@@ -634,7 +637,7 @@ static URI_INLINE int URI_FUNC(NormalizeSyntaxEngine)(URI_TYPE(Uri) * uri,
 				uri->hostText.afterLast = uri->hostData.ipFuture.afterLast;
 			} else if ((uri->hostText.first != NULL)
 					&& (uri->hostData.ip4 == NULL)) {
-				/* Regname and IPv6 */
+				/* Regname or IPv6 */
 				if (uri->owner) {
 					URI_FUNC(FixPercentEncodingInplace)(uri->hostText.first,
 							&(uri->hostText.afterLast));
@@ -649,7 +652,7 @@ static URI_INLINE int URI_FUNC(NormalizeSyntaxEngine)(URI_TYPE(Uri) * uri,
 					doneMask |= URI_NORMALIZE_HOST;
 				}
 
-				URI_FUNC(LowercaseInplaceAfterPercentEncoding)(uri->hostText.first,
+				URI_FUNC(LowercaseInplaceExceptPercentEncoding)(uri->hostText.first,
 						uri->hostText.afterLast);
 			}
 		}

ext/uri/uriparser/src/UriParse.c

@@ -196,7 +196,7 @@ static void URI_FUNC(OnExitPartHelperTwo)(URI_TYPE(ParserState) * state);
 
 static void URI_FUNC(ResetParserStateExceptUri)(URI_TYPE(ParserState) * state);
 
-UriBool URI_FUNC(PushPathSegment)(URI_TYPE(ParserState) * state,
+static UriBool URI_FUNC(PushPathSegment)(URI_TYPE(ParserState) * state,
 		const URI_CHAR * first, const URI_CHAR * afterLast,
 		UriMemoryManager * memory);
 
@@ -2119,7 +2119,7 @@ static URI_INLINE void URI_FUNC(ResetParserStateExceptUri)(URI_TYPE(ParserState)
 
 
 
-UriBool URI_FUNC(PushPathSegment)(
+static URI_INLINE UriBool URI_FUNC(PushPathSegment)(
 		URI_TYPE(ParserState) * state, const URI_CHAR * first,
 		const URI_CHAR * afterLast, UriMemoryManager * memory) {
 	URI_TYPE(PathSegment) * segment = memory->calloc(memory, 1, sizeof(URI_TYPE(PathSegment)));